Soggy Boots on Everest and Robotic Warfare: Why We Should Not Wait for the Arrival of Full Autonomy to Field Unmanned Ground Vehicles

Photo By Dan Heaton | Robotic Combat Vehicle information sheet, March 2021  

MI, UNITED STATES

03.18.2021

Courtesy Story

Next Generation Combat Vehicles Cross-Functional Team (NGCV CFT)

By: MAJ Cory Wallace, Next Generation Combat Vehicles Cross Functional Team
and
Mr. Kevin Mills, Combat Capabilities Development Command Ground Vehicle Systems Center


Before beginning any endeavor, one must ask themselves a question: Do I do this now or wait for better conditions?

The night prior to his historic summit of Mount Everest in 1953, Sir Edmund Hillary left his boots outside his tent. Naturally, the subzero temperatures froze them solid. Hillary and Tenzing, his Sherpa guide, did their best to thaw the boots using a camp stove and made the decision to push to the summit in less than ideal conditions. On 29 May 1953, Hillary and Tenzing became the first known humans to stand upon the highest point on Earth. The duo had the option of waiting for a better hand, but they elected to play the cards that the mountain dealt. Accordingly, history now binds their names to a monumental achievement as opposed to a footnote describing a failed attempt.

Now, we find ourselves facing a similar choice with regard to Robotic Combat Vehicles (RCV). We must decide if we should wait until unmanned ground vehicles (UGVs), such as the RCV, are able to autonomously maneuver through dense terrain, negotiate obstacles, and acquire targets for their human operators before we begin to integrate them into combat formations. Given what we have learned during the past four years of experimentation, we categorically believe that moving forward, while working with a still maturing level of technology, is the best path towards long-term success with bringing robotic warfare into fruition. Current technology will enable UGVs to provide significant value to both contemporary and future combat formations.

In 2020, the RAND Corporation conducted a company-level table top exercise (TTX) set in the Baltic States in the 2030s. Friendly forces (BLUFOR), consisting of a Manned Unmanned-Teamed (MUM-T) rifle company, augmented with both Light and Medium RCV variants and Unmanned Aerial Systems (UAS), attacked an enemy (OPFOR) rifle company defending a covered and concealed position. RAND conducted two iterations of the TTX. The first scenario, or the baseline, featured autonomy with a contemporary capability suite, to which we refer as “augmented teleoperation.” Further, the BLUFOR was only able to control RCVs if the control vehicles had a direct line of sight and in the absence of enemy electronic warfare attacks. Additionally, each control vehicle was limited to controlling two RCVs, thus “preventing [BLUEFOR] from massing RCVs against an [enemy] position” (RAND, 2020). Following the initial iteration, the BLUEFOR attacked with a very robust autonomy capability suite. During this second round, BLUEFOR RCVs could autonomously maneuver to positions of dominance without human intervention and converge lethal effects upon the enemy. These unmanned vehicles had “the ability to detect, identify, and engage targets without human intervention—which the Army does not yet envision—as an exemplar of the capabilities that might be technically feasible in the farther term” (RAND, 2020). In short, the autonomous vehicles “never failed to do what the players wanted” (RAND, 2020). Given the disparity between these capability sets, the results were little more than a forgone conclusion. In a Jan. 6, 2021, Forbes article, David Axe surmised the experiment’s results concluded that “remote-controlled vehicles are actually inferior to both manned vehicles and A.I. vehicles in certain key regards. Go with human beings or self-steering ‘bots, but maybe don’t try to comprise between the two” (Axe, 2021). Taking this single data point at face-value, one would logically have little, if any, desire to integrate UGVs into a combat formation prior to the arrival of full autonomy characteristics. We disagree with Axe’s conclusions.

Both virtual and live experiments have generated the Army’s initial robotic warfare concept. This vision includes UGVs and UASs operating forward of the human element and employing their sensors to rapidly develop a common operating picture that commanders will use to dictate the terms of the first human engagement. Unmanned vehicles will significantly extend the battlefield’s geometry and make first contact with an enemy well beyond the direct fire range of manned platforms. This form of ground maneuver warfare is a revolutionary change from our current doctrine and will require a substantial amount of time to learn how best to employ MUM-T on a hyperactive battlefield. The Army will build competence in this regard incrementally through ample “sets and reps.” Whether we begin this journey now or wait until 2040, the fact remains that robotic warfare is not something that will materialize overnight, regardless of autonomy’s maturity.

Integrating UGVs into combat formations prior to the arrival of full autonomy will allow Soldiers and leaders to develop and refine the necessary doctrine and tactics, techniques, and procedures (TTPs) required to effectively employ UGVs on future battlefields. Failing to couple the integration of new technology with validated doctrine and TTPs negates much of a new technology’s potential utility to the warfighter. For example, the US Army began using Night Vision Devices (NVDs) in the early years of Vietnam and continued to do so until the Army of the ‘80s and ‘90s boasted that it categorically “owned the night.” This capability gave the Army a substantial advantage over its adversaries up to the present day.

The Army built upon the lessons learned in Vietnam and continued to refine its doctrine and TTPs until it arrived at its current state of night-fighting proficiency. Speaking plainly, arriving at this point has taken a substantial amount of time and repetitions. Today’s Soldiers might scoff at the sight of a Starlight Scope or PVS-5s, but that archaic technology allowed their predecessors to own the night and led to today’s thermal night vision goggles and tomorrow’s Integrated Visual Augmentation System.

The same logic applies to the future revolutionary integration of UGVs into combat formations. Waiting for the perfect conditions that we presume will arrive upon the advent of full autonomy will deny years, if not decades, worth of training and doctrine development, thus offsetting a significant portion of UGV utility and effectiveness.

Our adversaries are certainly not waiting for perfect autonomy to move forward with integrating UGVs into their formations. Open source reporting provides one with a deluge of information about China and Russia’s efforts in this regard. In June 2020, The Daily Mail published both an article and footage of China’s “Sharp Claw 1” (Thomson, 2020). In a statement reminiscent to those made by our own Army leaders, Bai Mengchen explained that the PLA will have a “human-in-the-loop” to authorize lethal engagements and help the robot improvise and “halt [a] task when necessary” (Thomson, 2020). Meanwhile, Russia went as far as to test the Uran-9 platform in combat while conducting operations in Syria (Roblin, 2019). While the platform performed poorly during these operations, Russia took note of its shortcomings and is working to improve the system’s future performance to support Russia’s strategic objective of “deploying 30 percent of Russia’s kinetic weapons on remote-control platforms by 2025” (Roblin, 2019). Recall the example of the Army’s embrasure of NVDs during Vietnam which culminated with the dominance of operating in limited visibility conditions. Our adversaries are taking the same approach with UGVs. They are not waiting for the arrival of full autonomy for they see the value of an aggressive integration to optimize platform performance with mature doctrine and force structures. We cannot wait for perfect technological conditions while our adversaries enjoy a significant head start in the development and implementation of robotic warfare.

While maintaining pace with our adversaries is important, we must also develop a bond of trust between UGVs and their operators, Army leaders, and our civilian legislature. Humans are naturally apprehensive about embracing new technology that poses a potential of physical harm. For example, people refused to ride in elevators during the 19th century due to a fear that the system would fail, crash, and kill everyone foolish enough to ride in the deathtraps. This attitude changed with the inclusion of an elevator attendant who directed the elevator to the floor desired by its passengers. The human attendant helped establish a degree of trust that enabled the idea of an elevator transcend from a deathtrap to the common convenience that we know today (Hill, 2019). We must cultivate a similar degree of trust with UGVs if we hope to successfully integrate them into future combat formations.

In addition to building trust, integrating UGVs into combat formations will enable the Army to collect the data necessary to improve autonomy with relevant and reliable behaviors. Data collected from small-scale experiments and limited engineering evaluation tests will enable UGVs to perform well in those environments. But Soldiers and engineers employ combat vehicles in much different ways. Integrating UGVs into combat formations, maneuvering them in conjunction with a human element, and training them to perform tactical tasks in a relevant mission environment will provide future autonomy relevant to the warfighter. The Ground Vehicle System Center (GVSC) understands this approach and is placing it into practice with their Leader-Follower program. This effort consists of Palletized Load Systems (PLSs) autonomously following a human-operated PLS. Currently, Soldiers are operating a fleet of Leader-Follower vehicles at Fort Polk, Louisiana. Leader-Follower is not only influencing future autonomy, but this experiment is enabling developers to learn how to do the actual integration of unmanned systems into a human formation. Developers are also using the Soldier feedback to develop new TTPs and accurately define the value proposition of Leader Follower. To be specific, GVSC has learned that Leader-Follower has the potential to decrease the number of Soldiers required to provide logistic support and thus gives commanders options to re-task those Soldiers to other missions such as security augmentation or additional logistic resupply missions. The Army must take the same approach with UGVs and integrate them into combat formations to deliver the same results provided by the Leader-Follower program.

Let us further pull the thread of the actual definition of “full autonomy.” Both civilian and government efforts struggle to define what automated behaviors constitute this term. Everyone agrees that we must continue to work towards full autonomy; however, the moment at which we will reach our goal is still up for debate. Some viewpoints believe that full autonomy consists of behaviors that are almost as good as human behaviors while others demand that machines must bypass our capabilities before we can deem them fully autonomous. Anchoring the decision to move forward with integrating UGVs upon the arrival of full autonomy therefore hinges the entire effort upon a subjective opinion. If we believe full autonomy is ten years away, how can we be certain that we will be satisfied with what we find in 2031 or another arbitrary starting point? Goal posts tend to move when they are anchored on subjective opinions as opposed to tying decisions to objective performance measures such as a formation’s performance at a combat training center or the outputs of a Mission Essential Task List evaluation in a training environment. Both our live and virtual experiments have proven that UGVs enhance a formation’s lethality and survivability. We are where we need to be now. Let us move forward with the autonomy we have instead of waiting for better conditions that will always await on the far side of the next decade.

Moving forward with integrating UGVs into combat formations now will allow the Army to iterate upon the platform’s software and hardware requirements vis-à-vis Soldier and leader feedback. While there is a hardware technology element to combat UGVs, the driver for future capability, the elusive “full autonomy”, is a software defined future. Exquisite requirements defined in a “waterfall development process” are the antithesis of modern software development best practices. Software-defined UGVs offer the Army the ability to quickly iterate small, incremental capability improvements based on agile software sprints allowing tactics and technology to evolve together. Compare this approach vice moonshot requirements effort that could very well be focused on the wrong target. The specter of Future Combat Systems still haunts contemporary modernization efforts and serves as salient reminder of what happens when the Army develops requirements in a vacuum as opposed to embracing an iterative and dynamic process. Therefore, waiting for ideal conditions and hoping that we get the requirements right on the first attempt is yet another permutation of past failed endeavors.

A modern software developmental approach is not something the Army can simply “will” into existence; rather, the Army must take several critical steps to realizing this desire. The Army has dedicated significant investment into taking a modular open system approach (MOSA) for unmanned maneuver capabilities for UGVs. Leveraging the ubiquitous Robotic Operating System (ROS) middleware for autonomy, purposeful intellectual property management – combined with the new Software Acquisition authorities granted by Congress -- the Army has the ability to quickly and affordability improve baseline system capabilities driven by direct feedback from Soldiers using the systems. We acknowledge that implementing this vision will challenge every element of the Army Acquisition Enterprise, from how we write requirements to how systems are developed, acquired, tested, safety certified, and ultimately upgraded over their life cycle. That said, we categorically believe a modular and iterative approach is the path the gives the Army the best chance to get this critical bet right. This approach enables the Army to take advantage of innovation to get UGVs into the fight as quickly, efficiently and affordably as possible.

Let us now pose the counterargument and take the position that there is little point of moving forward with UGVs until unmanned vehicles are capable of autonomously flowing across battlefields in formations of RCVs paired with UASs that rapidly detect targets and strike our enemies miles behind their front lines. Anything short of that standard would waste both the Army’s time and taxpayer dollars. Let us also disregard this paper’s previously stated arguments. All things being equal, we still have one unmitigated challenge: Soldiers will continue to perform the most dangerous tasks on hyperactive and lethal future battlefields while we wait for full autonomy to arrive. Prior to that moment, Soldiers, not robots, will investigate potential chemical strikes, breach obstacles while under coordinated combined arms fire, and conduct reconnaissance on isolated and dangerous observation positions. These are all tasks that UGVs can perform, but our Soldiers will not be able to offload the extreme risk associated with these missions until the Army decides to integrate UGVs into combat formations. Integrating UGVs into combat formations is risky, but the potential consequences pale when compared to making a parlay bet to deliver both perfect doctrine, MUM-T force structures, and platforms on the first attempt after our adversaries have been experimenting with robotic formations for an entire decade.

In closing, let’s revisit our friend Sir Edmund Hillary shivering in a tent on Mount Everest in 1953. His boots were soggy and the aggregate conditions for summiting the highest point on Earth were far from ideal, but he and Tenzing took the risk and achieved their goal. Applying this metaphor to our argument, we and our adversaries are climbing towards employing MUM-T formations on future battlefields for we know the extreme lethality of the next conflict is something we have yet to fully appreciate. Successfully integrating UGVs into combat formations will take time, resources, and repetitions. Contemporary technology is adequate to enhance a ground maneuver formation’s effectiveness and will enable that formation’s capability to evolve as technology matures based on direct Soldier feedback. Our adversaries know this and have begun their own respective MUM-T journeys. And just as Hillary and Tenzing clawed their way to the top of Everest with soggy boots, we too must drive forward with our UGV effort while we enjoy relatively stable conditions in the world. Let us not wait for war to solve the problem. And should the veil of tranquility depart and we find ourselves at the precipice of a violent, engulfing global contact, the one question we do not want to ask ourselves is, “Why did we wait?”


Works Cited

Axe, David. (06 January 2021). “The U.S. Army’s Robot Tanks Will Make Great Bait.” Forbes. Retrieved 12 January 2021 from https://www.forbes.com/sites/davidaxe/2021/01/06/the-us-armys-robot-tanks-make-great-bait/?sh=5b50d3f54319.

Barnett, Sean D., Boston, Scott, Frelinger, David R., Gilmore, J. Michael, Gonzales, Daniel, Hou, Alexander C., Tarraf, Danielle C., and Whitehead, Peter. (2020). An Experiment in Tactical Wargaming with Platforms Enabled by Artificial Intelligence. RAND Corporation.

Hill, Rebecca A. (15 June 2019). “Can Robots Be Trusted?” Purdue Alumnus Magazine. Retrieved 12 January 2021 from https://www.purduealumnus.org/research/can-robots-be-trusted/.

Roblin, Sabastien. (06 January 2019). “Russia’s Uran-9 Robot Tank went to Syria (It Didn’t Go Very Well).” National Interest. Retrieved 12 January 2021 from https://nationalinterest.org/blog/buzz/russias-uran-9-robot-tank-went-war-syria-it-didnt-go-very-well-40677.

Thomson, Billie. (16 June 2020). “China Unveils ‘Small but Lethal’ War Robot ‘Sharp Claw I’ That’s Armed with a Machine Gun and Night Vision.” Daily Mail. Retrieved 12 January 2021 from https://www.dailymail.co.uk/news/article-8426027/Chinas-war-robot-Sharp-Claw-armed-machine-gun-night-vision.html.

• Date Taken: 03.18.2021
• Date Posted: 03.18.2021 11:49
• Story ID: 391719
• Location: MI, US